Google Pulls AI Image Tool From Google Earth After Misinformation Backlash

Google has abruptly disabled an AI image-generation feature in Google Earth just one day after its launch, following a wave of criticism and demonstrations showing how the tool could be used to create convincing misinformation on top of real satellite imagery.

The feature, powered by the company's Nano Banana 2 AI model, allowed users to type a text prompt to generate photorealistic images grounded in Google Earth's satellite, aerial and 3D imagery. Announced earlier this week and rolled out globally on the web version of Google Earth on Thursday, it was pitched as a way to visualize historical sites and real estate projects. But within hours, journalists and researchers had produced images of a collapsed Eiffel Tower, a sinkhole swallowing the Great Pyramid of Giza, and Russian tanks in Kyiv, raising alarms about the potential for abuse.

In a statement posted on X on Friday, Google said it had seen "people sharing screenshots of generated imagery that appear to violate our policies" and that it was "pausing this capability while we work on stronger guardrails." The company did not specify what types of images violated its policies, but said the images did not appear in the main Google Earth experience and were watermarked as AI-generated.

The decision came after BBC Verify tested the feature and was able to create a range of false scenarios, including a fake hospital in Gaza with a bomb crater next to it, a non-existent nuclear power plant in Iran, and a refugee camp on the U.S.-Mexico border. The images, shared by AI and misinformation expert Henk van Ess, who used the tool to generate them, underscored the ease with which realistic-looking fakes could be produced.

"The forgery does not have to look convincing on its own. It inherits the credibility of the map it was born on," van Ess told BBC Verify.

Google had initially defended the feature, saying it takes misinformation seriously and that every image created with Nano Banana in Google Earth includes the SynthID digital watermark, allowing users to check authenticity via the Gemini app or Google Lens. However, BBC Verify found that the watermark checks could be circumvented. In tests, the team was able to trick Gemini into saying that fake Google Earth images were real, and external AI detection tools also failed in some cases to identify the AI-generated content.

The feature also appeared to have inconsistent safeguards. BBC Verify found that while some prompts were rejected, small changes could bypass the guidelines. For example, a request to create a "raised platform to hang traitors" near the UK parliament was blocked, but a less specific prompt was accepted. Similarly, the tool refused to generate an image of the word "Trump" mowed into the White House lawn, but allowed a community garden in the same location.

Henry Ajdar, an AI-detection researcher, told BBC Verify that the creation of fake imagery in conflict zones could be particularly "destabilising," especially when events are fast-moving and information is scarce. "Journalists might be able to check that image... and report on it. But again, not everyone is going to do that," he said.

Geospatial analysts also warned about the broader implications for public trust. Bill Greer, a geospatial analyst and co-founder of the non-profit satellite service Common Space, noted that satellite imagery has historically been seen as an especially trustworthy source of data. "This is potentially particularly damaging to public trust," he told BBC Verify. He added that governments are increasingly restricting access to data, which could lead to more false images and fewer real ones, further eroding trust.

Google's quick reversal echoes a broader pattern among tech giants racing to integrate AI tools across their products, only to pull back when users abuse them. Earlier this month, Meta discontinued Muse Image, an AI feature that allowed users to generate images using public Instagram accounts, after widespread criticism over privacy concerns, including from a Hollywood union.

The incident highlights the tension between innovation and safety in the AI era. While Google has said it is working on stronger guardrails, the episode raises questions about the readiness of such tools for public use, especially when they are built on platforms that carry an inherent air of authority.

For now, the AI image-generation feature in Google Earth remains paused, and it is unclear when or if it will be reinstated. Google has not yet responded to questions about the specific policy violations that triggered the rollback, but the company's statement on X suggested that the pause is intended to allow time to strengthen safeguards.

As the use of AI-generated content becomes more widespread, the challenge of verifying what is real and what is fabricated is likely to grow. The Google Earth incident serves as a reminder that even well-intentioned tools can be exploited, and that the credibility of trusted sources, such as satellite imagery, is not immune to erosion.